Selection of parasitoids of the genus Trichogramma ( Hymenoptera : Trichogrammatidae ) and parasitism at different eggs ages of Duponchelia fovealis ( Lepidoptera : Crambidae )

The European pepper moth Duponchelia fovealis (Zeller, 1847) (Lepidoptera: Crambidae), despite being among the principal pests of strawberries in Brazil. However, biological control methods are rarely documented for this pest. In the current study, the goal was to select promising Trichogramma strains to enable the control of D. fovealis and to determine the acceptance of different ages of host eggs by the parasitoid. In the first bioassay, evaluations were done on the biological parameters of five strains of Trichogramma pretiosum Riley and one of Trichogramma galloi Zucchi (Hymenoptera: Trichogrammatidae). The best performance came from the strains, Tp18 (T. pretiosum) and Tg (T. galloi), showing parasitism and viability higher than 50 and 90%, respectively. Another bioassay revealed the acceptance of both these species (Tp18 and Tg) for different ages of D. fovealis eggs. All host ages were accepted by T. pretiosum (Tp18) and T. galloi, with a preference for younger eggs (65% parasitism). Adult emergence, number of individuals per egg and sex ratio revealed no significant differences with respect to the host age. The current study provides pertinent data on the selection and performance of Trichogramma species on D. fovealis eggs of different ages.

Although the potential for damage by the European pepper moth can be devastating, only a few studies have been done using alternative techniques to replace insecticides (Blok & Messelink, 2009;Efil et al., 2014;Messelink & Van Wensveen, 2003).As strawberry is consumed in vast quantities as fresh fruit, the application of insecticides poses a high risk of toxicity to humans, non-target organisms, as well as the environment (Fenik, Tankiewicz, & Biziuk, 2011), and the employment of the natural enemies in D. fovealis control is highly recommended as a safe management method (Zawadneak et al., 2016).
For egg parasitoids to be successfully used in biological control programs extensive knowledge of their biological features and host interactions is mandatory (Carvalho et al., 2014).Because Trichogramma is a cosmopolitan parasitoid, most of them exhibit interspecific and intraspecific variations in host preference (Hassan, 1989).However, the acceptance of certain hosts may be due to the search behavior and nutritional and morphological characteristics of the host egg (Farahani, Ashouri, Zibaee, Abroon, & Alford, 2016;Mansfield & Mills, 2004).One of the significant characteristics of the parasitoid-host interaction is the correct understanding of parasitoid acceptance when exposed to the host eggs of different ages (Pratissoli, Polanczyk, Pereira, Furtado, & Cocheto, 2007;Song et al., 2015;Zhang et al., 2014).The age of the host egg is crucial to its acceptance by the parasitoid, as advanced egg age can change the nutritional values and modify the parasitism rate (Farahani et al., 2016).
In the current study, the goal was to select promising Trichogramma strains to enable the control of D. fovealis and to determine the acceptance of Trichogramma pretiosum Riley and Trichogramma galloi Zucchi (Hymenoptera: Trichogrammatidae) at different ages of D. fovealis eggs.

Parasitoid and Host Rearing
Duponchelia fovealis, T. pretiosum and T. galloi are the species utilized in the bioassays drawn from the insects rearing of the Entomology Laboratory of the Universidade Federal do Espírito Santo (UFES-CCAE), in Alegre, Espírito Santo, Brazil.
Duponchelia fovealis was raised by placing freshly emerged adults in PVC tube cages (20 x 20 cm).At the base of each cage, a paper-coated styrofoam sheet was spread and a piece of voile fabric (30 x 30 cm) was placed at the cage top and fixed with a rubber strip.Sulfite paper was used to line the inside of the tube to facilitate the egg collection.Then, honey solution soaked in cotton was provided as food for the adults.The paper containing D. fovealis eggs was collected every day and disinfection with 0.5% formaldehyde solution and 17% copper sulfate was done for 10 seconds.The paper holding the eggs was then carefully packed in an acrylic box (11 x 11 x 3 cm) until the caterpillars emerged.These caterpillars were transferred to glass tubes (8.5 cm height x 2.5 cm diameter) and provided with the artificial diet according to King and Hartley (1985).Pupae were placed in pots having moist paper until the adults emerged.Duponchelia fovealis were laboratory reared at 25 ± 1°C, 70 ± 10% RH and 14 hours photoperiod.
Trichogramma species were reared on eggs of the alternative host Anagasta kuehniella Zeller (Lepidoptera: Pyralidae), according to Pratissoli et al. (2010).Host eggs were sterilized under a germicidal lamp for 50 minutes and glued to the surface of the cardboard pieces (8 x 2 cm) by a solution of "Arabic glue" (20%).Then, each cardboard piece was inserted into a glass tube (8.5 cm height x 2.5 cm diameter) and a newly emerged parasitoid was added to each tube, closed with PVC plastic film.

Selection of the Trichogramma strains
In the selection bioassay, six Trichogramma strains were assessed (Table 1).Two strains (Tp18 and Tg) were supplied by BUG Agentes Biológicos, São Paulo, Brazil, and the other four were indigenous strains (Tp 1, 8, 15 and 19) collected from different hosts and sites in the field (Table 1).After the morphological identification of the strains based on the protocol of Querino and Zucchi (2012), they were deposited in the Entomology Laboratory collection in the Universidade Federal do Espírito Santo (UFES-CCAE), Alegre, Espírito Santo, Brazil.From each Trichogramma strain, a female (age 0-5 hours) was selected and placed in individual glass tubes (8.5 x 2.5 cm).For each female, thirty D. fovealis eggs (up to 24 hours old) were glued on a cardboard piece (8 x 2 cm) using arabic glue solution (20%).The parasitism was allowed without interruption for 24 hours.After that, using a fine bristle brush, the females were removed from the tube, and the cardboards carrying the parasitized eggs were maintained in a climate chamber (25 ± 1°C, 70 ± 10% RH and 14 hours photoperiod).The variables percentage of parasitism of eggs, number of individuals per egg, emergence rate and sex ratio were assessed after ten days.

Effect of host egg age on the parasitism by Trichogramma strains
For this bioassay, the best strains of the previous bioassay (Tp18 and Tg) were used.The newly emerged T. pretiosum (Tp18) and T. galloi (Tg) females were isolated in glass tubes (8.5 x 2.5 cm).Each female was offered thirty D. fovealis eggs with 24, 48 and 72 hours, glued with arabic glue solution on one piece of cardboard (8 x 2 cm).A drop of honey was deposited on the inside of the tube wall to provide a carbohydrate source to the wasp.After 24 hours of parasitism, the parasitoids were removed.The cardboard pieces bearing the parasitized eggs remained undisturbed in the same glass tubes, which were once again sealed with PVC plastic film and placed in a climate chamber (25 ± 1°C, 70 ± 10% RH and 14 hours photoperiod).Once the parasitoids had emerged, the number of parasitized eggs, emergence rate, number of individuals per egg and sex ratio, were quantified.

Selection of the Trichogramma strains
Adopting the completely randomized experimental design, the choice of the Trichogramma strains was performed involving six strains with 15 replicates.The data were submitted to analysis of variance (ANOVA) and the means were compared by the Tukey test (p < 0.05), applying the Package ExpDes (Ferreira, Cavalcanti, & Nogueira, 2018) of the computer application R version 3.4.0(R Development Core Team, 2017).
To confirm which of the Trichogramma strains exhibited more similar characteristics, multivariate procedures were done.Using the averages of the variables of the strains in the study a 6 x 4 matrix (strains x explanatory variables) was built.First, this matrix was subjected to scalar and linear transformations of centralization and normalization to enable the Euclidean distance to be calculated later, and a matrix of dissimilarities 6 x 6 was obtained, in which the higher values between two lines were indicative of greater degree of distance, while the lower values implied greater degree of closeness (Borcard, Gillet, & Legendre, 2011;Mingoti, 2013).The dissimilarity matrix was thus utilized to identify the number of binding groups.To accomplish this, the Single Linkage method was employed to establish groups produced from closer (similar) strains.This procedure used 26 criteria (Charrad, Ghazzali, Boiteau, & Niknafs, 2014) to select and validate the number of binding groups for the Agglomerative Hierarchical Methods (Borcard et al., 2011;Mingoti, 2013).From the number of groups thus defined, the dendrogram was constructed using the matrix of dissimilarities.Therefore, to understand the effect of the explanatory variables on cluster formation, one principal coordinate analysis (PCoA) was performed using the matrix of dissimilarities.Subsequently, a "Biplot" graph was drawn to interpret the behavior of the variables (Borcard et al., 2011).
All the analyses were performed utilizing the computational application R version 3.4.0(R Development Core Team, 2017).For the multivariate analysis procedures, the vegan package was employed (Oksanen et al., 2017), and the number of clusters was determined by applying the NbClust package (Charrad et al., 2014).Using the completely randomized design, the host age bioassay was performed in a 2 x 3 factorial scheme, with two Trichogramma species, three egg ages (24, 48 and 72 hours) and ten replicates.Data were submitted to analysis of variance (ANOVA).In the case of significant interaction, the interaction between the factors "species" and "egg ages of D. fovealis" was performed, with the means of each combination (treatments) compared by Tukey's test (p < 0.05).However, if the interaction was not significant, the factors "species" and "ages of eggs" would be analyzed in isolation.Thus, as the factor "species" has only two levels would be the decisive ANOVA analysis (F test).As for the factor "age of eggs", because it has three levels, the average would be compared by Tukey test (p < 0.05) utilizing the computational application R version 3.4.0(R Development Core Team, 2017).

Selection of the Trichogramma strains
All the Trichogramma strains tested parasitized the D. fovealis eggs.The Tg (15.36) and Tp18 (15.30) strains revealed the highest value of parasitized eggs (F 5, 84 = 13.548;p < 0.001; Table 2), which corresponded to a parasitism above 50%.The emergence of parasitoids in the treatments range between 92.93 to 97.57%, showing no significant difference (F 5, 84 = 2.061; p = 0.0783; Table 2).In the case of number of the individuals per host egg, no significant difference was observed among the strains tested (F 5, 84 = 2.207; p = 0.0610), showing values close to one individual per host egg (Table 2).A statistical difference among the means of the sex ratio was noted for the Trichogramma strains (F 5, 84 = 6.075; p < 0.0001).Sex ratio values higher than 0.9 were observed, except in the case of the Tp1 strain, which exhibited the lowest sex ratio (0.79) (Table 2).
From the multivariate analysis the clear evidence of four groups emerged, based on the dissimilarity matrix and Single Linkage method (Figure 1A).The dendrogram constructed showed that the groups thus formed included the following strains: group 1 -Tp15; group 2 -Tp19 and Tp8; group 3 -Tp1; and group 4 -Tg and Tp18 (Figure 1B).
The PCoA analysis revealed five explanatory dimensions, in which the first two dimensions were consistent with 81.76% of the variation observed (Figure 2A).The PCoA indicates the placement of the groups and explanatory variables (Figure 2B).
On analysis of the first dimension (greater percentage of variation) group 4 (Tg and Tp18) was confirmed to have the highest parasitism value.Groups 2 (Tp19 and Tp8) and 3 (Tp1), however, showed lower parasitism values.Group 1 (Tp15) revealed an intermediate parasitism value, although it was placed near to group 4. For the first dimension, the variables number of individuals per egg, emergence and sex ratio did not exhibit significant influence.However, for the second dimension, the effect of the variables parasitism, number of individuals per egg and sex ratio, was clearly evident, which enabled the distinction between groups 3 and 4, respectively, to be made.The distinction between groups 1 and 2 resulted from the combined interpretation of both dimensions, as these groups exhibited closely similar characteristics, with the strong expression of the variable parasitism, in the grouping.

Effect of host egg age on the parasitism by Trichogramma strains
No significant interaction between the factors of the "Trichogramma species" and "egg age" was observed for any variable studied: parasitized eggs (F 2, 54 = 1.011, p = 0.370), number of individuals per egg (F 2, 54 = 0.157, p = 0.854) and emergence (F 2, 54 = 0.659, p = 0.521).The sex ratio also did not show variation, since all the progenies were female, therefore it was not submitted to ANOVA.In this way, the factors were analyzed separately.

Discussion
In this study, a selection of the Trichogramma strains on the D. fovealis eggs was first performed.All strains parasitized D. fovealis eggs, with higher outcomes for the Tp18 and Tg strains.Later on, the best strains (Tp18 and Tg) were employed in a bioassay to the parasitism at the different egg ages of D. fovealis.It was evident that the Tp18 and Tg strains parasitized the different egg ages of D. fovealis, showing a clear preference for the younger eggs.
Four of the six strains assessed in this study are indigenous (Tp 1, 8, 15 and 19) while two are commercial (Tp18 and Tg).From the multivariate analyses performed, Tp18 and Tg were identified as the most effective strains on D. fovealis, which corresponded to the T. pretiosum and T. galloi species.Normally, these commercial strains are subjected to several selection and quality control tests, and possess the potential to overpower populations of diverse host species (Lenteren, 2003).However, the findings regarding the indigenous strains in this study appear to be promising.
The indigenous strains, Tp15 and Tp1, exhibited parasitism of 12.92 and 10.87 eggs, respectively, in 24 hours of host exposure.Parasitism would probably increase if the time of exposure to the host was higher, because parasitoids of the genus Trichogramma can normally parasitize 10 or more days, based on the prevailing environmental status (Carvalho et al., 2014;Pratissoli, Bueno, Bueno, Zanúncio, & Polanczyk, 2009).In addition, indigenous strains can tolerate much better adverse climatic conditions and thus be more effective against their hosts (Herz & Hassan, 2006).
The results showed that the age of D. fovealis eggs significantly influenced the number of eggs parasitized by the T. pretiosum (Tp18) and T. galloi (Tg).Higher parasitism was noted on the 24 and 48 hours eggs, with no significant difference between the 48 and 72 hours eggs.Besides, parasitism was observed also in more than 50% of the D. fovealis eggs presented to the parasitoids, regardless of age of the host egg.However, the decline in parasitism observed among the treatments in both the Trichogramma species highlights the preference exhibited by the parasitoid for the D. fovealis eggs in the initial phase of embryonic development.Other studies have also that the older host egg can reduce Trichogramma parasitism in many host species, such as Chilo suppressalis (Walker) (Lepidoptera: Crambidae) (Ko et al., 2014;Zhang et al., 2014), Bonagota salubricola (Meyrick) (Lepidoptera: Tortricidae) (Pastori, Monteiro, Botton, & Pratissoli, 2010) and Plutella xylostella (L.) (Lepidoptera: Plutellidae) (Pratissoli et al., 2007).Such preference for younger eggs may result from an alteration in the egg reserve nutrients.As the embryonic development advances, the egg reserve nutrients are transformed into chemically more complex tissues which render them less appealing to the parasitoid (Vinson, 1997).Another factor that may have exerted some influence on the number of parasitized eggs is the chorion hardening in the older eggs, which may limit its acceptance by the parasitoid, posing a hindrance in penetrating the ovipositor of the parasitoid (Mellini, 1987).
In terms of the viability of the parasitized eggs, an emergence rate above 95% was observed for the different ages of the D. fovealis eggs.Therefore, even in the more advanced embryonic developmental phases, D. fovealis eggs were suitable for the development of T. pretiosum and T. galloi.The high viability rates, over 85%, are acceptable for biological control programs (Almeida, Silva, & Medeiros, 1998).Besides, in this study the emergence rate recorded were higher than those in other studies, such as T. galloi on eggs of Diatraea saccharalis Fabricius (Lepidoptera: Crambidae) (Oliveira, Santana, Bellon, & Oliveira, 2014) and T. pretiosum on the eggs of B. salubricola (Pastori et al., 2010) with an emergence rate below 50 and 60%, respectively, for the eggs of different ages.Therefore, the findings of the high emergence rate in this study indicate the potential of T. pretiosum and T. galloi as possible and effective control agents of D. fovealis.This factor is significant for success in the establishment of inoculative biological control programs, as even in the event of high parasitism, the low emergence rate can influence the biological control outcomes (Bueno, Parra, Bueno, & Haddad, 2009).
Trichogramma pretiosum and T. galloi, did not present significant differences in the number of individuals per host egg, however, with values near to one individual per egg.The D. fovealis eggs were observed to supply adequate nutritional support irrespective of egg age.To arrive at the number of one individual per egg can be considered a good finding for the establishment of biological control programs for D. fovealis.The emergence rate of more than one individual per host egg can give rise to smaller sized and poorer quality parasitoids, which can affect the parasitoid performance caused by intraspecific competition (DaSilva, Morelli, & Parra, 2016;Moreira, Santos, Beserra, Torres, & Almeida, 2009).
Trichogramma pretiosum and T. galloi species produce female individuals only as the descendants.This emergence rate of solely females for both the Trichogramma species may be linked to the favorable developmental conditions provided by the host, because factors like age and host quality, size of the postures and superparasitism affect the development and therefore the sexual ratio (Vinson, 1997).Besides, the progeny of this parasitoid may result from the behavioral decisions taken by the females during host acceptance, thus enabling the parasitoids to optimize the sexual ratio of the offspring to suit the local conditions (Luck, Janssen, Pinto, & Oatman, 2001).However, other factors that could influence the sex ratio in the Trichogramma populations, include the presence of the bacterium Wolbachia, which produces reproductive alterations in many arthropods, inducing thelytokous parthenogenesis (Boivin et al., 2014) and the presence of an extra chromosome called PSR (Paternal Sexual Reason) or the selfish genetic element, which can convert an egg destined to be a diploid female into a haploid male with PSR (Russell & Stouthamer, 2010).
The European pepper moth D. fovealis is one of the principal strawberry pests in Brazil.At present, there is neither record of any plant protection products available in Brazil to control or manage this insect, nor any effective measures of biological control.Of the Trichogramma species assessed in this study, T. pretiosum strain Tp 18 and T. galloi showed great promise, revealing high rates of parasitism, emergence rate and production of female offspring, which are the fundamental characteristics to selection of parasitoids for biological control programs.Besides, the two Trichogramma species parasitized D. fovealis eggs of all ages, showing a sharp preference for the younger eggs.

Figure 1 .
Figure 1.Number of clusters based on 26 criteria for group selection (A) and cluster analysis (B) of the different Trichogramma strains using Euclidean distance and the Single Linkage method.Strains belonging to the same group are the same rectangle.

Figure 2 .
Figure 2. Percentage of explanation of the dimensions of the principal coordinate analysis (PCoA) of the Trichogramma strains (A) and "Biplot" graph (B) showing the effect of the explanatory variables on the clusters formed.Groups of lineages: 1: gray, 2: green, 3: blue and 4: red.

Table 1 .
Trichogramma strains, hosts, crops and collection site/company.Effect of host egg age on the parasitism by Trichogramma strains
Means followed by the same letter do not differ statistically by the Tukey test (p > 0.05).* Variables are statistically the same by the F test (p > 0.05).

Table 3 .
Biological parameters (±SE) of Trichogramma pretiosum (Tp18) and Trichogramma galloi (Tg) reared on Duponchelia fovealis eggs of different ages.Temperature 25 ± 1°C, RH 70 ± 10% and photoperiod 14 hours.Means followed by same uppercase letter for interaction do not differ by Tukey's test (p < 0.05).*Means followed by same lowercase letter for factors do not differ by Tukey's test (p < 0.05).ns Not significant by F test (ANOVA) (p >0.05).NS Not significant by Tukey test (p > 0.05). 1 It was not analyzed; all values were equal.